Highway System Damage from Pers 15Aug2007 M8.

0 Earthquake
Jerome S. OCONNOR, P.E., MCEER, University at Buffalo, Buffalo, NY; Lucero E. MESA, P.E. South Carolina Department of Transportation; and Monique NYKAMP, P.E., Shannon & Wilson, Inc., Seattle, Washington
IBC-08-99
KEYWORDS: Post-Earthquake, Damage Assessment, Seismic, Reconnaissance, Bridges, Highways, Liquefaction, Pisco, Peru ABSTRACT: The performance of Pers highway system in this M8.0 earthquake is relevant since the infrastructure was built using AASHTO specifications. The Pisco earthquake of July 2007 was a strong and unusually long-period earthquake. Soil conditions and ground water levels led to widespread liquefaction contributing to more highway damage than ground shaking did. This reconnaissance survey illustrates why geotechnical considerations are paramount. Because some bridges were subject to secondary hazards such as rockfall, there is also a case for consideration of multiple hazards during the design or retrofit of bridges. INTRODUCTION On August 15, 2007, a magnitude Mw 8.0 earthquake struck the vicinity of Pisco, Per. In September, MCEER conducted a reconnaissance mission to investigate and document damage to the highway system. The trip was organized jointly with an American Society of Civil Engineers (ASCE)s lifeline field reconnaissance team, led by Alex Tang, P.E. The authors comprised the highway system investigation team and provide this paper as an overview of the event and the consequential damage to roads and bridges. Because of the need to restore critical infrastructure as quickly as possible, some repair was underway or already completed when the reconnaissance team arrived in September. The investigators met with Pers Ministry of Transportation (MTC) and were provided firsthand accounts and official reports from affected regions that were written in the hours and days after the event. The authors findings include these assessments, information gathered from leaders in Pers academia who had done field investigations, as well as their own personal field observations. The purpose of the report is to document the performance of structures designed according to AASHTO specifications and to help assess the adequacy of the standards used at the time of construction. Another important benefit of the task is the exchange of information for the mutual benefit of engineers and policy makers here in the U.S. and in Per. Per is in a very seismically active region and documentation of performance during events such as the Pisco earthquake can yield benefits when considering modifications to the design specifications in our country. The investigation took place September 21-28, 2007. Both U.S. Customary (English) and SI (metric) units are used in this report. Per uses SI units, but U.S. readers of this report may be more comfortable with English units. The MCEER report (3) that this paper summarizes contains an English-Spanish translation of technical terms that will be useful to English speakers who wish to better their understanding of engineering terms used in Per and other Spanish speaking countries. It has terms that are pertinent to transportation engineering, bridge inspection, and geotechnical engineering. BACKGROUND As seen in figure 1, the epicenter of the earthquake was just off the west coast of South America. Per has a population of 28 million, most of who live in the capital Lima. The earthquake was felt in Lima but damage was concentrated to the south in the Ica region where approximately 700,000 reside. In this coastal region, there are arid deserts and sand dunes, with rainfall in the

region amounting to less than 1/2 per year. This contrasts greatly from the eastern part of the country, where there is rocky terrain and abundant water from Andean mountain streams. Despite this arid climate, the region suffers from occasional flooding when El Nio rains fall in the highlands. These rains result in a phenomenon known as huayco (or huaico), which is a violent debris flow that can present a hazard to structures. Ica was devastated by flooding of the Ica River in 1963, then again, in 1983 and 1998, when rains from El Nio resulted in inundation of 80% of the city. Soils near the coast consist largely of liquefiable silts and sand. Where firm ground exists near Lima, it is an alluvial deposit locally known as Lima conglomerate, a dense, slightly cemented gravel. In the Ica/Pisco area, fine soils are cemented eolian deposits. Away from the coast, the ground is very rocky with little arable soil. Pers highway system consists of three categories of roads: National, Departmental (regional), and Rural (lightly traveled unpaved roads). The backbone of the National Highway System (NHS) is the Pan-American Highway (Carretera Panamericana Sur), which is the countrys primary North-South route. All NHS roads are run by concession, meaning that the national government has a long-term (30-year) contract with a private company that serves as an operator and collects tolls in exchange for maintaining the roads in good condition. In the event of an earthquake, the contractor has a contractual obligation to restore the highway system to its original condition. Per relies on AASHTO specifications for the design of its bridges. The 475-year event found in the Standard Specification for Highway Bridges has typically been used. Additional precautions must have been incorporated into the designs because the team observed bridges with wide bridge seats and transverse restraints such as concrete shear blocks. Piles or caissons are not typically used for foundations, apparently due to lack of necessary equipment. In Per, only structures exceeding 10 meters are classified as bridges. The building code in Per divides the country into three seismic zones for both building and bridge design. Peak ground accelerations (PGA) are specified for each zone with the highest, 0.40g, being used for the coastal zone. THE EARTHQUAKE The so-called Pisco earthquake of August 15, 2007 (6:41 pm local time) was caused by a subduction of the oceanic

Nazca Plate under the South American Continental Plate. It resulted in at least 519 deaths, 1,874 injuries, and tens of thousands of buildings destroyed or severely damaged. Per has a long history of earthquakes. In 1650, a strong earthquake destroyed the city of Cusco, which was the capital of the Incan empire that stretched from what is now Chile to Colombia. In 1664, an earthquake devastated the city of Ica. As recently as 2001, an Mw 8.4 earthquake struck off the coast in the same region as this earthquake.

Figure 1 Map of South America and Per with Epicenter of the August 15, 2007 Pisco Earthquake Peak ground accelerations (PGAs) as high as 0.49 g (approximately half the acceleration due to gravity) were recorded on accelerometers at the PCN station in Parcona, Ica. Time history graphs of the earthquake show two distinct phases of strong ground motion and an unusually long duration, resulting in shaking that lasted over 170 seconds. GEOTECHNICAL EFFECTS The primary geotechnical impacts of the earthquake were due to the phenomenon of liquefaction. Liquefaction occurs during ground shaking in loose, saturated, sandy soil when the water pressure in the pore spaces increases to a level that is sufficient to separate the soil grains from each other. This phenomenon results in a reduction of the shear strength of the soil (a quicksand-like condition). Liquefaction can result in ground settlement, lateral

spreading, landsliding, localized ground disruptions from sand boils (ejection of sand and water at the ground surface), and reduced vertical and lateral capacity for structure foundations. Buildings, bridges, and other structures founded on or in the liquefied soils may settle, tilt, move laterally, or collapse. Because much of the affected area consists of fine grained, water-saturated sediments, the geotechnical aspects of the earthquake were most dramatic. There was evidence of widespread liquefaction, especially along the coastal regions where the water table was not far from the surface. Newspapers had accounts of water spouting 2 m (6.5) up from the ground during the earthquake in Ica. Even near Lima where the PGAs were less than 0.1 g, there were utility poles tipped from liquefaction of the supporting soils. Figure 2 provides an interesting crosssection of a sand boil. This phenomenon results when pore pressure in lower layers of soil force increase to the point that liquefied soil up through the surface.

earthquakes unusually long duration (>170 seconds). Liquefaction likely caused one major slope failure on the Pan-American Highway that led to 75 mm (3) wide shear cracks in a three-cell concrete box culvert. It also caused some parts of the roadway to shift laterally and the pavement to be broken up and faulted. It most likely precipitated damage to one abutment of the Huamani Bridge that was observed. This five-span bridge also had shear cracking of piers and a 100 mm (4) lateral movement of the superstructure. Fifteen significant rockfalls blocked roads, but most travel ways were opened again by the time of the field investigation. There were at least two instances of failed retaining walls observed by the team. Falling boulders hit at least two large bridges causing serious damage. Approximately twenty-four sites were inspected. This includes bridge structures and other highway features such as culverts, roadways, embankments, and retaining walls. The team discovered damage at 62% of these sites. The MTC was not able to provide an estimated cost of damage to the highway system and the investigators did not try to quantify the value of the damaged facilities. PAN-AMERICAN HIGHWAY - On September 25, 2007, the highway system investigation team drove south from Lima to the Department (state) of Ica, where earthquake damage was most evident. In the Ica region, it is a twolane road, paved with asphalt. Locations along the PanAmerican Highway are conveniently posted as the number of kilometers (km) south of the capital Lima. The Pan-American Highway is parallel to the coastline, and as such, is built upon sandy and silty soils. The terrain is coastal desert with large sand dunes and sparse vegetation. The Pacific Ocean is as close as 100 m (328) to the west. The road is bordered on the east by large dunes consisting of cemented sand formations. At numerous stops along the highway, the team encountered evidence of earthquake-induced liquefaction and lateral spreading. Lateral spreading occurs on gentle slopes as a result of soil liquefaction. When the soil liquefies, gravity causes the land to move downslope. Conversations with individuals that had conducted inspections immediately following the earthquake recounted stories from locals stating that the water table was high prior to the event. Acceleration records of the event show that there were ground vibrations for an extended period (over 170 seconds). This ground motion, soil amplification and liquefaction resulted in damage to the pavement, embankment and cut slopes, drainage structures and shoulders, as evident from the numerous photographs in the MCEER report.(3)

Figure 2 Geologist Dr. Carlos Costa, UNSL in Argentina pointing out soil layers within cross section of sand boil during paleoliquefaction study. He was working with Patricio Valderrama, Per Geological Survey. ROADWAYS Although there was widespread destruction of buildings in the Ica region of Per, damage to the highway system was less severe. Traffic on the Pan-American Highway, which is the backbone of the highway system, was interrupted at numerous points but most repairs were made within a few weeks. The one exception was the Huamani Bridge, which was still closed for repair six weeks after the earthquake. Highway infrastructure suffered some damage from shaking, but most damage resulted from earthquake induced liquefaction, which was accentuated by the

almost dry. Flooding, however, does occasionally occur. The El Nio effect, which last occurred in 1998, has left a very wide riverbed as its legacy.

Figure 3 Embankment and roadway failure at km 190. The picture was taken facing north. Note the Pacific Ocean in the top left of the photo. The damage was caused by amplification of the ground shaking. Liquefaction of the wet coastal silts and sands led to lateral spreading and the embankment failure.

Figure 5 Pan-American Highway near San Clemente at km 220.

Figure 4 Pan-American Highway near San Clemete at km 220. DEPARTMENTAL ROUTE 100 This is a secondary road on the Ica Departmental Road Network (see figures 6 and 7). It starts at the Pan-American Highway in Chincha and heads in an easterly direction. This is an unpaved road with no appurtenances such as signage or guiderailing. Km posts denote the distance away from Chincha. As one travels east, the terrain becomes more and more rugged. The road winds around the rocky foothills of the Andes mountains and generally follows the San Juan River, which is also known as the Chincha River. The river carries water down from the Andes mountains towards the ocean but most of the time the volume of water is very low. By the time it gets to the ocean, the water seeps underground and the river is

Figure 6 Monique Nykamp and remains of several rockfalls that blocked Route 100 at km 29. STRUCTURE PERFORMANCE PUENTE HUAMANI - Huamani Bridge is a significant link of the highway system that was taken out of service by the earthquake. It is located at km 224 of the PanAmerican Highway and is located directly east of the epicenter. This concrete structure was built in 1950 to cross the Rio Pisco (Pisco River) to join the city of Pisco with the village of San Clemente to the north. It is 136 meters (446) long and two lanes wide (6.7 m or 22). The bridge was designed to an H-15 AASHTO loading and had been posted with a 36-ton weight limit.

Severe cracking of one pier stem Severe cracking of transverse concrete diaphragms Minor to severe cracking and spalling of horizontal shear blocks and pier caps Tipping of the south abutment toward the stream bed (approximately 100 mm (4) at the top) Lateral displacement of the superstructure (approximately 100 mm or 4) Damage to approach pavement, especially at south end of the bridge

Figure 7 Road damage from failed retaining wall on Route 100 at km 27. Although all five spans of the bridge remain standing, there was evidence of liquefaction in the sandy soils and evidence of strong shaking. The abutments and piers are built on massive walls, similar to caissons but are not supported by piles. According to record plans obtained by Ing. Jack Lpez Jara, the 4 m (13.1) high pier stems rest on solid concrete caissons that are founded 7.8 m (25.6) into the stream bed. There was no indication of tipping or settlement of the piers but the south abutment tipped toward the stream because of liquefaction of the sandy soil at that end of the bridge. During the inspection, the superstructure was being temporarily supported and the south abutment was being rehabilitated. The north abutment is on a more competent material (a lightweight siltstone) that did not liquefy. There was only hairline cracking of that abutment. There was cracking of the top of the pier caps, concrete diaphragms and shear blocks, apparently resulting from lateral movement of the superstructure. The second pier from the south end had particularly severe damage from this lateral loading. See figure 9. The fact that the superstructure was only moderately displaced is evidence of the successful use of shear blocks. Although the bridge was closed and extensive repair was necessary, the bridge did not collapse. Immediately after the earthquake, the bridge was reopened with traffic restricted to just one lane. Shortly thereafter, however, it was closed to all traffic as a precautionary measure and so that repair work could be accomplished. The bridge was under repair at the time of the inspection. Bridge damage can be summarized as:

Had the bridge not been part of a lifeline transportation route, this performance could be considered acceptable. The bridge did not collapse and was successful in preventing loss of life. The damage it suffered, though severe, is repairable. Prior to the earthquake, the government had already been planning to build a new bridge at this site within the next few years.

Figure 8 The Huamani Bridge was still closed for repair five weeks after the earthquake. TIMBER BRIDGES Several bridges with log superstructures were inspected. There was minor cracking observed in the cheek walls from lateral forces associated with the earthquake but they performed surprisingly well. With the superstructure being relatively light (i.e., not much mass), the effect of ground accelerations were not so great. The bridges could be seen deflecting under heavy truckloads in this rural area and it is much more likely that they would fail from overload than an earthquake.

Figure 9 Severe cracking of Pier 2 (from South end) necessitates extensive repair but the horizontal shear blocks managed to retrain lateral movement of the superstructure. This pier has roller expansion bearings.

Figure 11 Puente Huachinga elevation. Note its precarious position in the shadow of a large mountain from which rocks fall.

Figure 10 Approach pavement at the south abutment of Huamani Bridge has settled dramatically due to liquefaction of supporting soils and lateral spread of the embankment. PUENTE HUACHINGA This steel truss, built in 1966, is vulnerable to rockfalls and huaycos as well as earthquakes. PUENTE LOS MOLINOS - The Los Molinos Bridge crosses the Ica River. It is a five span, one lane bridge that was built in 1932. The superstructure is structurally non-redundant, consisting of two cast-in-place concrete girders.

Figure 12 Puente Huachinga at km 39. This steel truss bridge has suffered severe damage to the bottom chord from presumably previous debris impact (top arrow). There is a large granite boulder jammed between the two channels of the bottom chord that has fallen from the adjacent mountain from this earthquake (bottom arrow). The bridge may have been more vulnerable to earthquake damage because of unrepaired damage stemming from the flooding that accompanied the 1998 El Nio. Scour at Pier No. 1 exceeds two meters and Pier No. 2 is tipped. There is a noticeable kink in the alignment of the bridge, partially because of the tipped pier but possibly because of a lateral shift from the recent earthquake. According to a local resident, it was like that before, but worse now.

Figure 14 Puente Los Molinos. Note the pre-existing scour (>2m or 6) at Pier No. 1 (as identified from left to right) (see arrow) and tipped Pier No. 2 shown with the arrow on the right. SURVIVORS The City of Ica had strong ground shaking. Although the team saw numerous adobe buildings that had failed, the bridges inspected did not show any evidence of damage. The bridges were cast-inplace concrete built within the past twenty years, so the design and type of construction probably played an important part in their good performance.

Figure 13 Puente Huachinga at km 39. It is uncertain whether this bottom chord damage was caused by falling boulders from this event or whether it remains from a previous hazard, such as El Nio effect flooding that last occurred in 1998. This deformation can limit the capacity of this tension member and put additional stress in the undamaged channel. At three different locations, there are shear cracks that appear to be new. Although they could be a result of truck loading, they appear to be from a lateral stress to the girders applied at the supports. Although there was not excessive displacement, this type of damage is consistent with earthquake induced lateral loading. Evidence of liquefaction was noted at this bridge by several sand boils. Liquefaction of the alluvial soils in the riverbed may have also contributed to observed damage to the bridge. Liquefaction likely resulted in reduced lateral support to the bridge piers. Although the bridge could be considered fragile because of the previous scour and the newer cracking of the structural members, it is a vital link in the transportation system. During the inspection, several fully loaded trucks crossed the bridge hauling out rubble from nearby towns.

Figure15 This example of newer bridge construction performed well. CONCLUSIONS Much of the Pan-American Highway, which tracks the coastline, is built on liquefiable soil and suffered damage. Extensive sections of the highway subsided and needed to be rebuilt. Considering the intensity and duration of shaking, damage to bridges was minimal. The Huamani Bridge suffered the greatest damage and was still closed for repair six weeks after the event.

Observed damage included: Rockfalls at fifteen or more locations. Although most roads had been cleared of rock debris, there were still numerous locations where fractured, unstable blocks of rock above the roadway pose a hazard. Liquefaction of sandy/silty soils leading to tipping of utility poles, and tipping of a bridge abutment at Puente Huamani. It also initiated lateral spread that displaced an entire section of roadway, sheared a three-cell concrete box culvert, caused an embankment slope failure, and severe cracking and faulting of asphalt pavement. Two or more were bridges damaged by boulders falling from an adjacent). Cracking of concrete pier caps and shear blocks (Puente Huamani). The shear blocks, designed to provide lateral resistance to the bridge superstructure, successfully restrained the bridge so that damage is repairable. Shear cracking of concrete girders at supports (Puente Los Molinos). This appears to have been from lateral movement of the superstructure. Lateral displacement of bridge superstructures (about 100 mm or 4) (Puente Huamani and Puente Los Molinos). REFERENCES 1. Edwards, C. L. (Ed.) (2002). Atico, Peru Mw 8.4 Earthquake of June 23, 2001, American Society of Civil Engineers, 176 pages, ISBN: 0-78440661-8. Edwards, C., Yashinsky, M., and Byers, W. (2003). Highways and Railroads, paper from Southern Peru Earthquake of 23 June 2001 Reconnaissance Report, Earthquake Spectra, Supplement A to Volume 19, January 2003, pages 97-114. OConnor, J. S., Mesa, L., Nykamp, M. (2007). Damage to the Highway System from the Pisco, Peru Earthquake of August 15, 2007, Technical Report MCEER-07-0021, December 10, 2007. Wartman, J., Rodriguez-Marek, A., Repetto, P. C., Keefer, D.K., Rondinel, E., Zegarra-Pellane, J., and Baures, D. (2003). Ground Failure, paper from Southern Peru Earthquake of 23 June 2001 Reconnaissance Report, Earthquake

Spectra, Supplement A to Volume 19, January 2003, pages 35-56. ACKNOWLEDGMENTS The post-earthquake investigation was funded by FHWA under Project DTFH61-98-C-00094, which is led by MCEER Special Tasks Director George C. Lee, Ph.D., under the direction of Phil Yen, Ph.D. Special thanks go to Jack Lpez Acua, Jack Lpez Ingenieros S.A.C. and Jack Lpez Jara, T.Y. Lin International, Inc. in Lima Per for their technical and logistical assistance. Photo credits by figure numbers: 3 by Ing. Jack Lpez Jara; 4 & 9-Dr. Ing. Jorge E Alva Hurtado; 5 & 10 by Dr. Julio Kuroiwa; 7 by Ing. Lucero Mesa. All other photos by Jerome OConnor.

2.

3.

4.